Prof. P. N Suganthan Nanyang Technological University Title: Randomization Based Deep and Shallow Learning Methods for Classification Keynote Day: 19 December 2020 Abstract: This talk will first introduce the origins of randomization-based feedforward neural networks such as the popular instantiation called random vector functional link neural network (RVFL) originated in early 1990s. Subsequently, a performance comparison among the randomization-based feedforward neural network models will be presented. The talk will also include ensemble/deep randomization-based neural networks. Another randomization-based paradigm is the random forest which exhibits highly competitive performances. The talk will briefly describe heterogeneous oblique random forest. Kernel ridge regression will be also briefly introduced. The talk will also present extensive benchmarking studies using classification datasets. Bio: Ponnuthurai Nagaratnam Suganthan received the B. A degree, Postgraduate Certificate and M. A degree in Electrical and Information Engineering from the University of Cambridge, UK in 1990, 1992 and 1994, respectively.After completing his PhD research in 1995, he served as a pre-doctoral Research Assistant in the Dept. of Electrical Engineering, University of Sydney in 1995–96 and a lecturer in the Dept. of Computer Science and Electrical Engineering, University of Queensland in 1996–99. He moved to Singapore in 1999. He was an Editorial Board Member of the Evolutionary Computation Journal, MIT Press (2013-2018) and an associate editor of the IEEE Trans on Cybernetics (2012-2018). He is an associate editor of Applied Soft Computing (Elsevier, 2018-), Neurocomputing (Elsevier, 2018-), IEEE Trans on Evolutionary Computation (2005-), Information Sciences (Elsevier, 2009-), Pattern Recognition(Elsevier, 2001-) and Int. J. of Swarm Intelligence Research (2009-) Journals. He is a founding co-editor-in-chief of Swarm and Evolutionary Computation (2010-), an SCI Indexed Elsevier Journal. His co-authored SaDE paper (published in April 2009) won the "IEEE Trans. on Evolutionary Computation outstanding paper award" in 2012. His former PhD student, Dr Jane Jing Liang, won the IEEE CIS Outstanding PhD dissertation award, in 2014. IEEE CIS Singapore Chapter won the best chapter award in Singapore in 2014 for its achievements in 2013 under his leadership. His research interests include swarm and evolutionary algorithms, pattern recognition, forecasting, randomized neural networks, deep learning and applications of swarm, evolutionary & machine learning algorithms. His publications have been well cited (Googlescholar Citations:~39k). His SCI indexed publications attracted over 1000 SCI citations in a calendar year since 2013. He was selected as one of the highly cited researchers by Thomson Reuters every year from 2015 to 2018 in computer science. He served as the General Chair of the IEEE SSCI 2013. He is an IEEE CIS distinguished lecturer (DLP) in 2018-2020. He has been a member of the IEEE (S'91, M'92, SM'00, Fellow’15) since 1991 and an elected AdCommember of the IEEE Computational Intelligence Society (CIS) in 2014-2016. Address: S2-B2a-21, EEE, NTU, Singapore, 639798 Tel: 65-67905404 Fax: 65-67933318
Prof. Yihong Gong Xi’an Jiaotong University University Title: Brain-Inspired Machine Learning Methods Keynote Day: 19 December 2020 Abstract: In this talk, I will present four novel methods that are inspired by human brain visual cognitive mechanisms. First, we propose the Min-Max objective function inspired by the manifold separation property of human visual cortex, which enforces the CNN model to learn features with minimized within-class distances and maximized between-class distances. Second, we propose the L-21 norm-based objective function inspired by properties of neurons in the V-4 layer of human ventral pathway. It enforces the sparse category selectivity on neurons in the output layer of a CNN model. Third, we propose the CNN structure that is inspired by the dual-pathway mechanism of the human visual system and is able to solve the “texture bias” problem of the existing CNN models. To solve the “catastropic forgetting” problem that occurs when fine-tuning a CNN model with new training samples, we propose the Anchor Loss objective function that requires the CNN model to keep the topological structure of the learned feature space during the fine-tuning phase. This work is inspired by the latest cognitive scientific research on human visual memory. We also developed several methods to automatically capture topological structure of a learned feature space.These proposed objective functions are independent of and can be applied to any CNN models. Comprehensive performance evaluations show remarkable performance improvements of the representative CNN models on the respective tasks without increasing their model complexities. Bio: Yihong Gong is a distinguished professor, the dean of School of Software Engineering of Xi’an Jiaotong University, an IEEE Fellow, and a vice director of the National Engineering Laboratory for Visual Information Processing. His research interests include image/video content analysis and machine learning algorithms. He is among the first batch of researchers in the world initiating research studies on content-based image retrieval, sports video event detection, text/video content summarization, and image classification using the sparse coding image features. He has published more than 200 technical papers and two monographs. To date, his works have received more than 22,000 citations (Google h-index=64), with over 3,500 citations for his most cited paper. In 2015, his ACM SIGIR 2003 paper titled “Document Clustering Based on Non-Negative Matrix Factorization” received “Test of Time Award” Honorable Mentions by the ACM SIGIR Executive Committee. Under his supervision, his teams have won numerous international/domestic competitions in image/video content recognitions.
Prof. Qi Wang Northwestern Polytechnical University Title: Crowd Counting Research towards Real World Application Keynote Day: 19 December 2020 Abstract: In recent years, due to the frequent occurrence of large-scale activities, crowd density estimation is becoming significant. With the continuous development of deep learning and computer vision, the performance of crowd counting methods has been greatly improved. This report will introduce the crowd counting research towards the real world application. It mainly includes three aspects: (1) utilize virtual data to build a large-scale annotated crowd counting data set, and improve the generalization ability of crowd counting models through supervised learning and domain adaptation; (2) propose an inter-domain feature isolation model to translate synthetic data into real data, and use Gaussian priors to improve the quality of the density map addressing the problem of inter-domain differences and the generation of fine crowd density maps; (3) establish a large crowd counting data set and a benchmark for researchers to evaluate the algorithm performance, which will promote the rapid development of the crowd counting. Bio: Qi Wang received the B.E. degree in automation and the Ph.D. degree in pattern recognition and intelligent systems from the University of Science and Technology of China, Hefei, China, in 2005 and 2010, respectively. He is currently a Professor with the School of Computer Science and with the Center for OPTical IMagery Analysis and Learning (OPTIMAL), Northwestern Polytechnical University, Xi’an, China. His research interests include computer vision and pattern recognition. He is the Section Editor-in-Chief of Remote Sensing, and the associate editors of IEEE T-CSVT, IEEE T-SMC:System, IEEE GRSL, etc.
|
Prof. Shigang Yue University of Lincoln Title: Dealing with motion in the dynamic visual world – from insects to neuromorphic sensors Keynote Day:19 December 2020 Abstract: Animals, as small as insects, have amazing ability in coping with the dynamic visual world. This ability has not been replicated so far to human-made intelligent moving machines such as robots or intelligent vehicles. In the last decades, different types of visual neurons in insects have been identified with surprising preferences tuned to specific visual cues. By modelling bio-plausible visual neurons and their pre-synaptic networks, we can not only further our understanding of how animals visual systems work, but also step forward in developing new vision systems for future robots and vehicles. In this talk, I will introduce the bio-inspired motion sensitive neural models developed in my group, and also talk about their applications in robotics and autonomous vehicles. Bio: Shigang Yue is a Professor in the School of Computer Science , University of Lincoln, United Kingdom. He is the founding director of Computational Intelligence Lab (CIL) and the deputy director of Lincoln Centre for Autonomous Systems (L-CAS). He received his PhD degrees from Beijing University of Technology (BJUT) in 1996, worked in BJUT as a Lecturer (1996-1998) and an Associate Professor (1998-1999), also in City University of Hong Kong (MEEM) as a Senior Research Assistant (1998-1999). He was an Alexander von Humboldt Research Fellow (2000, 2001) working with Prof. Henrich in the Faculty of Computer Science, University of Kaiserslautern, Germany. Before joining the University of Lincoln as a Senior Lecturer (2007) and promoted to Professor (2012), he held positions in the University of Cambridge (2006-2007), University of Newcastle (2003-2006) and the University College London (UCL) (2002-2003) respectively.
His research interests are mainly within the field of bio-inspired artificial intelligence, computer vision, robotics, brains, and neuroscience. He is particularly interested in biological plausible visual neural systems and its applications in vehicles, interactive systems, UAVs and ground robots. He has published about 200 papers in academic journals and conferences, many of them are in top tier journals. He has chaired several international conferences and has sit in the editorial board of several international journals. Prof. Shigang Yue home pages and short cv: http://webpages.lincoln.ac.uk/syue http://www.ciluk.org/syue https://staff.lincoln.ac.uk/syue
|
